It is known that modern internal combustion engines are provided with a fuel injection system for directly injecting the fuel into the cylinders of the engine. As an example, the so called Common Rail System (CRS) is the most used one for Diesel Engines. The CRS, generally, comprises a fuel pump, hydraulically connected to a fuel common rail and a plurality of electrically controlled fuel injectors, which are individually located in a respective cylinder of the engine and which are hydraulically connected to the fuel rail through dedicated injection pipes.
As also known, the injection pressure is one of the most important parameter determining the quality of the fuel injection into the engine (for example, the fuel spray penetration in the cylinder head) and must be regulated as function of the engine working conditions, for example, according to a map engine load vs. engine speed. A known technique used to control the injection pressure is a feed forward technique coupled with a proportional and integral (PI) closed loop control.
A problem of the actual pressure injection control and all controls using a standard feed forward technique is that the full compensation via feed-forward is not practicable, and therefore this implies the key role of the PI integrator as the only way to compensate the system unknowns and, consequently, to null the regulation error. On the other side: the system unknowns require a large span for the PI integrative and any saturation applied to the integrator output for anti-windup purposes should allow the regulation error cancellation; the need to filter the feed-forward makes heavier the PI integrative workload during transients: for a given slew-rate, stronger the filtering is higher the response overshoot is; the anti-windup technique could be ineffective during strong transients with high slew-rate: this could imply limitation on set-point dynamic and integrative gain.
Therefore a need exists for a new feed forward strategy that could overcome the above problems, let the PI integrator only acting on the compensation of system unknowns and tolerance. Therefore, at least one object is to provide a method of automatic control using a new feed forward technique, which coupled with a PI closed loop control, let to de-link the integrative gain KI from the overshoot requirements. At least another object is to provide an apparatus that allows performing the above method. In addition, other objects, desirable features and characteristics will become apparent from the subsequent summary and detailed description, and the appended claims, taken in conjunction with the accompanying drawings and this background.